limenghao opened a new pull request #7021: URL: https://github.com/apache/tvm/pull/7021
Add features in the paper (AdaTune: Adaptive Tensor Program Compilation Made Efficient). Main changes: 1. In `./python/tvm/autotvm/measure/measure_methods.py`. Add an optional argument `enable_adaptive_evaluator` in the measure_option, when it is set as true, it will use the adaptive evaluator to do the evaluation. 2. Add RFEITuner and RFEICostModel in `tvm/autotvm/tuner/`. Add an optional argument `uncertainty_aware` in the ModelBasedTuner. When it is set true, it will use an dynamic epsilon to control the epsilon-greedy searching process. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
